Abstract
We report a study examining the effect of implicit affirmation and denial in reasoning with negative conditionals. Four conditional inference problems are constructed by an affirmation of the antecedent (Modus Ponens, MP), an affirmation of the consequent (AC), a denial of the antecedent (DA) or a denial of the consequent (Modus Tollens, MT). Depending on the presence of negations in the conditionals, these problems were set with an implicit or explicit type of referencing (i.e. affirmation or denial). Implicit affirmation and denial introduced matching bias effects: When the topic of the categorical premise (e.g., the letter B) does not match the topic of the conditional's referred clause (e.g., If there is a letter A), participants made fewer conditional inferences. The effect of implicit affirmation (e.g., ‘the letter B’ affirms an antecedent ‘if there is no A’) was larger than the effect of implicit denial, and reading times (obtained by eye movement recordings) showed that implicit affirmation problems took longer to solve whereas implicit denial problems were solved faster than the explicit denial problems. Both findings are novel and we hypothesize that the larger effect on implicit affirmation of a negative (as compared to implicit denial of an affirmative) is due to uncertainty in considering a single instance sufficient to affirm the entire contrast class of that negative.